Modern wireless telecommunications systems are evolving to provide high speed packet data services for users of mobile equipment. One example is an ability to provide internet access to a user of mobile equipment. A wireless system that is rapidly evolving in this direction is a Time Division, Multiple Access (TDMA) system known as the Global System for Mobile Communication (GSM), in particular enhanced versions of GSM known as GSM+ and EGPRS (Enhanced General Packet Radio Services).
As can be appreciated, an important consideration when it is desired to provide high speed data communications in a wireless system is the various radio path signal impairments that can occur, such as fading and noise. If not properly accounted for, the presence of these impairments can result in the occurrence of unacceptable or objectionable transmission/reception error rates.
More particularly, for EGPRS it has been recently proposed that link quality measurements, namely mean Bit Error Probability (BEP) and a coefficient of variation (cv) (BEP), should be individually filtered for a sequence of measurement data in the ME before being reported on the uplink to the wireless network (ETSI SMG2 Tdoc 1048/99: “EGPRS Link Quality Control Measurements and Filtering”, Ericsson, 20-24 Sep. 1999, Bordeaux, France.) In this document it was proposed to use a running average exponential filter that is characterized by its “forgetting factor” (or averaging period) being broadcast to all of the ME in the serving cell, the same as in the General Packet Radio Service (GPRS). The forgetting factor was proposed to be broadcast using a pre-existing Packet Broadcast Control Channel (PBCCH), if available, otherwise a Broadcast Control Channel (BCCH) would be used. In essence, this approach would provide a cell-based forgetting factor.
Subsequently, it was proposed in a document ETSI SMG2 Working Session on EDGE #11 Tdoc 2E99-501:“EGPRS LQC Measurements filtering”, Nokia, 18-22 Oct. 1999, Austin Tex., USA, to replace the above-referenced forgetting factor by one that follows the definition for GPRS. It was pointed out that the filter proposed in the ETSI SMG2 Tdoc 1048/99 was not stable when the number of filtered data is below the so-called filter length. In order to counteract the filter instability the GPRS-based forgetting factor was proposed to be used.
This approach may be referred to as a non-fixed cell-based approach, in that it is similar to that found in GSM 05.08 v. 6.3.0, section 10.2.3.2 (Derivation of Channel Quality Report). In this approach, when the number of measurement data in the ME is less than the filter length, the forgetting factor is progressively adjusted so as to minimize the filter non-stability. It was also proposed to broadcast a parameter on the PBCCH, if available, otherwise on the BCCH, which is used by the ME in the serving cell to calculate the filter parameter, in combination with the amount of measurement data held by the ME. However, all ME having a sufficient amount of measurement data (exceeding the length of the filter) will use the same filter parameter.
FIGS. 1 and 2 are taken from the ETSI SMG2 Working Session on EDGE #11 Tdoc 2E99-501 document, and are graphs depicting weighting coefficients using various proposed forgetting factors. For a large forgetting factor (e.g., a=0.9) the measurement report made by the ME is dominated by the latest few measurement data, whereas when using a small forgetting factor the filtered results relate instead to a large number of past measurement data. By example, for the case of a=0.3, the filtered results correspond to over 100 past measurement blocks (about two seconds of measurement data.)
It was noted in the ETSI SMG2 Working Session on EDGE #11 Tdoc 2E99-501 document that within one cell various MEs might experience different channel conditions/fading effects (slow fading, fast fading, etc.) As a result, the viability of a filter with a cell-based forgetting factor (fixed or non-fixed) was questioned. Although not specifically addressed in this document, it was further noted that it would be of interest to study further whether a solution using a fixed cell-based forgetting factor (or GPRS-like) could, optionally, be replaced by more optimal one.
In summary, currently deployed or proposed systems use a fixed/non-fixed cell-based filter for each mobile equipment (ME) in the serving cell. However, these solutions are less than optimal for all of the MEs that may be present within the cell. That is, they are at most a compromise, as it is well known that different MEs within the same cell will encounter different channel conditions and fading effects, at least some of which can be directly linked to the velocity of the MEs. As it is most improbable that all MEs within a given serving cell will be moving with the same velocity, any cell-based filter that makes such an assumption is bound to provide less than optimal results.